Better guard ergonomics

[meh]

I've started using guards in my codebase and it would be nice if they could be a little more flexible.

Things I found a bit annoying so far:

1. Taking a path instead of an ident so I can fully/partially scope guards instead of having to import them.
2. Having to provide empty parentheses when the guard does not take any parameter.

What I'm going for is something like this:

#[graphql(guard(guard::Session))]

Instead of:

#[graphql(guard(SessionGuard()))]

Thanks!

[meh]

Oh also having to import async_graphql::guard::Guard for the proc macro to work is kind of annoying.

[sunli829]

I've started using guards in my codebase and it would be nice if they could be a little more flexible.

Things I found a bit annoying so far:

1. Taking a path instead of an ident so I can fully/partially scope guards instead of having to import them.
2. Having to provide empty parentheses when the guard does not take any parameter.

What I'm going for is something like this:

#[graphql(guard(guard::Session))]

Instead of:

#[graphql(guard(SessionGuard()))]

Thanks!

Because of the limitation of procedural macros, I haven't thought of a better way to remove this annoying empty parentheses. 🙂

[meh]

Ah, damn, that's understandable then, what about supporting paths instead of just an ident tho? 🐼

[sunli829]

It should work, I am trying.

[sunli829]

I have implemented it on the master branch. 🙂

[meh]

You're the best! Thanks!